Gadd45b Loss Ablates Innate Immunosuppression in Cancer Daniela Verzella1, Jason Bennett2, Mariafausta Fischietti1, Anil K

Published OnlineFirst December 26, 2017; DOI: 10.1158/0008-5472.CAN-17-1833

Cancer Tumor Biology and Immunology Research

GADD45b Loss Ablates Innate Immunosuppression in Cancer Daniela Verzella1, Jason Bennett2, Mariafausta Fischietti1, Anil K. Thotakura2, Camilla Recordati3, Fabio Pasqualini4, Daria Capece1, Davide Vecchiotti1, Daniel D'Andrea2, Barbara Di Francesco1, Marcella De Maglie3, Federica Begalli2, Laura Tornatore2, Salvatore Papa2,5, Toby Lawrence6, Stuart J. Forbes7, Antonio Sica4,8, Edoardo Alesse1, Francesca Zazzeroni1, and Guido Franzoso2

Abstract

T-cell exclusion from the tumor microenvironment (TME) is a basis to interpret clinical evidence that elevated expression of major barrier to overcoming immune escape. Here, we identify a GADD45B confers poor clinical outcomes in most human cancers. myeloid-intrinsic mechanism governed by the NF-kB effector Furthermore, they suggest a therapeutic target in GADD45b for molecule GADD45b that restricts tumor-associated inflammation reprogramming TAM to overcome immunosuppression and and T-cell trafficking into tumors. In various models of solid T-cell exclusion from the TME. cancers refractory to immunotherapies, including hepatocellular Significance: These findings define a myeloid-based immune carcinoma and ovarian adenocarcinoma, Gadd45b inhibition in checkpoint that restricts T-cell trafficking into tumors, with myeloid cells restored activation of proinflammatory tumor- potentially important therapeutic implications to generally associated macrophages (TAM) and intratumoral immune infil- improve the efficacy of cancer immunotherapy. Cancer Res; 78(5); tration, thereby diminishing oncogenesis. Our results provide a 1275–92. 2017 AACR.

Introduction have the effect of promoting oncogenesis (1, 2). Accordingly, tumor-associated inflammation is now considered a hallmark of Virtually all tumors contain an inflammatory infiltrate of innate cancer (1). The ability of the immune system to influence onco- and adaptive immune cells (1). Despite their inherent capacity to genesis is currently being exploited to treat cancer patients, with the counter neoplastic progression and eliminate nascent tumors, development of many successful anticancer immunotherapies (3). these cells, especially those of the innate immune system, often Indeed, therapies blocking pivotal immunoinhibitory mechanisms, comprising immune checkpoint molecules, for example, cytotoxic T lymphocyte antigen-4 (CTLA-4) or programmed cell 1Department of Biotechnological and Applied Clinical Sciences, University of death protein 1 (PD-1), are revolutionizing the clinical manage- L'Aquila, L'Aquila, Italy. 2Centre for Cell Signalling and Inflammation, Depart- ment of certain malignancies, such as metastatic melanoma (4). ment of Medicine, Imperial College London, London, United Kingdom. 3Mouse & However, major challenges remain, as the majority of patients 4 Animal Pathology Laboratory, Fondazione Filarete, Milan, Italy. Department of rarely exhibit an objective response to these treatments, due to the Inflammation and Immunology, Humanitas Clinical and Research Center, Roz- 5 absence of a preexisting intratumoral T-cell infiltrate or, in patients zano, Milan, Italy. Current address: Leeds Institute of Cancer and Pathology – fi (LICAP), University of Leeds, Leeds, United Kingdom. 6Centre d'Immunologie de with T-cell in ltrated tumors, the dominant inhibitory effects of Marseille-Luminy, Aix Marseille Universite, Inserm, CNRS, Marseille, France. additional tumor microenvironment (TME)–associated molecules 7Medical Research Council Centre for Regenerative Medicine, University of (3, 5). Yet, the mechanisms underlying the immunoinhibitory Edinburgh, Edinburgh, United Kingdom. 8Department of Pharmaceutical activities of the TME are presently poorly understood. Sciences, Universita del Piemonte Orientale 'Amedeo Avogadro', Novara, Italy Tumor-associated macrophages (TAM) are implicated in var- Note: Supplementary data for this article are available at Cancer Research ious TME-mediated mechanisms that enable cancers to evade Online (http://cancerres.aacrjournals.org/). immune attack (3, 6, 7). TAMs are the main leukocyte population D. Verzella, J. Bennett, M. Fischietti, and A.K. Thotakura are co-first authors of found in most human tumors, where they play a key role in this article. restricting local T-cell trafficking and T-cell effector functions F. Zazzeroni and G. Franzoso are co-last authors of this article. (3, 6, 7). Indeed, a high density of TAMs, especially those exhibit- ing an anti-inflammatory and immunosuppressive phenotype, Corresponding Authors: Guido Franzoso, Imperial College London, Room 10N8, Commonwealth Building, Hammersmith Campus, Du Cane Road, London, W12 correlates with poor clinical outcome in most human cancers 0NN, United Kingdom. Phone: 44-0-20-3313-8421; Fax: 44-0-20-83832788; E-mail: (7, 8). Accordingly, therapeutic approaches limiting myeloid cell [email protected]; Francesca Zazzeroni, Department of Biotechnological recruitment into tumors have resulted in increased TME-based þ and Applied Clinical Sciences, University of L'Aquila, Via Vetoio 10 – Coppito II, CD8 effector T-cell infiltration and diminished tumor burden L'Aquila 67100, Italy. Phone: 39-0862-433526; Fax: 39-0862-433523; E-mail: in mouse models and early-phase clinical trials (3, 6, 9–13). [email protected] However, the general macrophage depletion associated with doi: 10.1158/0008-5472.CAN-17-1833 these approaches raises important safety concerns. Therefore, 2017 American Association for Cancer Research. a preferable approach would be to reprogram TAMs toward

www.aacrjournals.org 1275

Verzella et al.

a proinflammatory phenotype capable of redirecting T-cell traf- cancers that are largely refractory to immunotherapies, including ficking into tumors and unleashing local antitumor immune hepatocellular carcinoma (HCC) and ovarian adenocarcinoma, responses (6, 14). we showed that Gadd45b deletion in myeloid cells restores þ Macrophages can differentiate into a spectrum of phenotypic proinflammatory TAM activation and intratumoral CD8 states in response to diverse environmental signals. At sites of T-lymphocyte infiltration, resulting in diminished tumor growth. infection, "classically" activated or proinflammatory macrophages Because we previously showed that GADD45b additionally med- (hereafter also occasionally referred to as M1-like macrophages) iates the NF-kB antiapoptotic activity in cancer cells (24), our arise in response to IFNg and Toll-like receptor (TLR) ligands to current findings identify GADD45b as a pivotal downstream hub eliminate invading pathogens, promote inflammation, and integrating the NF-kB oncogenic functions linking cancer and engage the adaptive immune system. In contrast, "alternatively" inflammation. Our finding that elevated GADD45B expression activated or anti-inflammatory macrophages (hereafter also occa- correlates with poor clinical outcomes across most human cancers sionally referred to as M2-like macrophages) are "educated" by IL4 consolidates the general clinical significance of the GADD45b- and IL13 at sites of injury to terminate inflammation and enable mediated oncogenic mechanism in malignant disease. Together, wound healing (7, 8). Interestingly, anti-inflammatory macro- these results reveal a pathogenically critical, innate immunity phages resemble the TAMs found in most human cancers, where "checkpoint" governed by GADD45b that is amenable to thera- they support tumor progression, metastatic dissemination, and peutic intervention to "re-educate" TAMs and ultimately over- cancer immune evasion (7, 8). However, little is known about come TME-dependent immunosuppression, with profound the mechanisms governing the opposing functions of macro- implications for anticancer therapy. phages in response to tissue injury, malignancy, or infection. Consequently, there are presently no therapeutic agents capable of effectively reprogramming TAMs to oppose oncogenesis. Materials and Methods Recent studies have identified NF-kB transcription factors as Human cancer datasets key determinants in the balance between proinflammatory and The human datasets of lung cancer (LUNG), stomach adeno- anti-inflammatory macrophage activation in response to infec- carcinoma (STAD), nonalcoholic liver hepatocellular carcinoma tion or tumor-derived signals (15, 16). In fact, in addition to (LIHC), esophageal carcinoma (ESCA), cervical carcinoma orchestrating protective immune and inflammatory responses, (CESC), untreated primary glioblastoma multiforme (GBM), the NF-kB pathway drives oncogenesis in multiple cancer types by cholangiocarcinoma (CHOL), head and neck squamous cell suppressing apoptosis of tumor cells and, concurrently, governing carcinoma (HNSC), and kidney clear cell carcinoma (KIRC) were TME-based inflammation, thereby serving as a central hub linking part of The Cancer Genome Atlas (25) program and downloaded cancer and inflammation (16, 17). In ovarian adenocarcinoma from the UCSC Cancer Genomic Browser (26). Gene expression models, macrophage-specific inhibition of the NF-kB–activating profiling was performed on fresh or frozen tissue biopsies using kinase, IkBa kinase (IKK)b, was shown to reverse the anti- the Illumina HiSeq 2000 RNA Sequencing platform. The esti- inflammatory TAM phenotype and enhance proinflammatory mates of GADD45B expression levels were derived from the macrophage activation, thereby causing tumor regression (18). normalized values in the UCSC Cancer Genomic Browser. The A similar conversion from an anti-inflammatory TAM activa- datasets of colon adenocarcinoma (COAD; GSE39582), bladder tion state supportive of oncogenesis to a proinflammatory carcinoma (BLCA; GSE13507) and ovarian cancer (OV; activation state that antagonizes tumor growth was reported GSE9891) were deposited in the National Center for Biotechnol- in melanoma and fibrosarcoma models in mice bearing TAMs ogy Information (NCBI) Gene Expression Omnibus (GEO) data- lacking NF-kB1/p105 (19). Notably, this IKKb/NF-kBfunction base (www.ncbi.nlm.nih.gov/geo). The COAD dataset was from in suppressing proinflammatory activation was shown to the French national Cartes d'Identite des Tumeurs (CIT) program extend beyond the cancer context, reflecting the tissue-specific and generated using the Affymetrix Human Genome U133 Plus role that IKKb/NF-kB plays in myeloid cells, where opposite to 2.0 Array platform (27). The BLCA dataset was from the Chung- its conventional proinflammatory role, NF-kB drives the reso- buk National University Hospital (Cheongju, South Korea) and lution of inflammation to enable wound healing and prevent generated using the Illumina human-6 v2.0 expression beadchip tissue damage (20, 21). platform (28). The OV dataset was from the Australian Ovarian Therefore, the IKKb/NF-kB pathway provides an attractive Cancer Study, Royal Brisbane Hospital (Queensland, Australia), route to therapeutically reverse TME-mediated immunosuppres- Westmead Hospital (Sydney, Australia), and Netherlands Cancer sion by reprogramming TAMs toward a proinflammatory phe- Institute (Amsterdam, Netherlands) and generated using the notype (15, 18). However, therapeutically targeting this pathway Affymetrix Human Genome U133 Plus 2.0 Array platform with conventional IKKb/NF-kB inhibitors has not proven possi- (29). The relative GADD45B mRNA expression levels in these ble, due to the severe on-target toxicities associated with globally datasets were derived from the normalized values present in the suppressing NF-kB (17). A logical alternative to pharmacologi- GEO database. The breast carcinoma dataset (BRCA) was from cally inhibiting IKKb/NF-kB would be therefore to target the the tumor banks in the United Kingdom and Canada, and nonredundant downstream effectors of the NF-kB anti-inflam- gene profiling data were generated using the Illumina matory function in myelomonocytic cells. However, these effec- HumanHT-12 V3 platform and deposited on Oncomine Research tors and their modes of action remain poorly understood. Premium Edition (30). Here, we report that the NF-kB–regulated protein, GADD45b The gene profiling data from each dataset were downloaded (22, 23), mediates an essential myeloid-intrinsic mechanism together with the accompanying clinical information. Where governing proinflammatory macrophage activation and the possible, the series were obtained from patients at an early disease þ immunosuppressive activity of the TME that restrict CD8 T-cell stage and datasets with a sufficient number of patients with trafficking into tumors. Using three distinct models of solid recurrence-free survival (RFS) information. In the other cases,

1276 Cancer Res; 78(5) March 1, 2018 Cancer Research

Breaking Cancer Immunosuppression via GADD45b Deletion

where this was not possible, the series consisted of overall survival Mice (OS) data and/or the entire patient dataset. Patients were stratified C57BL/6Jx129/SvJ Gadd45b / mice were described previously þ þ into two groups on the basis of the GADD45B mRNA expression (32). Gadd45b / littermates were cohoused and used as controls. þ þ levels. In each case, quintiles, quartiles, tertiles, and 95th C57BL/6J Gadd45b / and Gadd45b / mice were obtained by percentiles were used as thresholds, and the best fits are reported backcrossing the corresponding C57BL/6Jx129/SvJ lines onto in Fig. 1A–M. C57BL/6J to at least the N14 backcross generation. C57BL/6J þ CD45.1/Ly5.1 mice (B6.SJL-Ptprca Pepcb/BoyJ) and C57BL/6J Cell culture and macrophage isolation and treatment LysM-cre transgenic mice were purchased from The Jackson þ þ The C57BL/6 mouse fibrosarcoma cell line, MCA-203, was Laboratory. C57BL/6J Gadd45b / and Gadd45b / mice, expres- þ þ kindly provided by I. Marigo and V. Bronte (Department of sing either CD45.2/Ly5.2 or CD45.1/Ly5.1 , were cohoused and Oncology and Surgical Sciences, University of Padua, Padua, bred as separate colonies. Italy) and cultured in high-glucose DMEM (with L-glutamine, Conditional Gadd45b knockout mice, carrying loxP-flanked without sodium pyruvate; Gibco) supplemented with 10% Gadd45b alleles (Gadd45bF), were generated at Taconic Bios- heat-inactivated FBS (Sigma-Aldrich), antibiotics (150 U/mL ciences using conventional gene targeting technology in embry- penicillin, 200 U/mL streptomycin), 10 mmol/L HEPES onic stem (ES) cells. Homozygous Gadd45bF/F mice were obtained F/þ DM/DM (Gibco), 2 mmol/L b-mercaptoethanol, and 2 mmol/L L-gluta- by the interbreeding of Gadd45b mice. Gadd45b mice, mine (Gibco). The C57BL/6 mouse ovarian carcinoma cell line, specifically lacking Gadd45b in the myeloid lineage, were gener- ID8-Luc, stably expressing firefly luciferase (Luc), was described ated by crossing Gadd45bF/F mice with C57BL/6J LysM-cre mice previously (18) and cultured in high-glucose DMEM (with L- (The Jackson Laboratory), expressing a Cre recombinase transgene glutamine, without sodium pyruvate; Gibco) supplemented under the control of the myeloid-specific LysM promoter (33). with 4% heat-inactivated FBS (Sigma-Aldrich), antibiotics Cohoused Gadd45bF/F littermates were used as controls for (150 U/mL penicillin, 200 U/mL streptomycin), ITSþ1Liquid the experiments. Media Supplement (5 mg/mL insulin, 5 mg/mL transferrin, 5 ng/ Mice were housed in ventilated cages in a pathogen-free mouse mL sodium selenite; Sigma-Aldrich), and 2 mmol/L L-gluta- facility of the Central Biomedical Services at Imperial College mine (Gibco). Cells were cultured in a humidified incubator in London (London, United KIngdom) and used in accordance with 5% CO2 at 37 C. All cell lines were routinely tested using a established institutional guidelines, under the authority of UK Mycoplasma Detection Kit (ATCC). Before injection in vivo,cell Home Office project license, P75F16A53 [see Guidelines on the lines were also routinely screened for other infectious agents Operation of Animals (Scientific Procedures) Act of 1986]. Sep- using the Mouse Essential CLEAR Panel (Charles River Labo- arate mouse colonies were maintained in filtered top cages on ratories). The cell line authentication for the MCA-203 and ID8- autoclaved food, water, and bedding in the Conventional Mouse Luc murine cell lines has not been possible because these cell Facility at the University of L'Aquila (L'Aquila, Italy), and exper- lines are not commercially available. Cells were used for the imental procedures were performed in accordance with national experiments between 5 and 7 days upon thawing. and international laws and regulations (see European Economic Bone marrow–derived macrophages (BMDM) were prepared Community Council Directive 86/609, OJ L 358, 1, December 12, þ þ from 6- to 10-week old Gadd45b / and Gadd45b / C57BL/6J 1987; Italian Legislative Decree 116/92, Gazzetta Ufficiale della mice as described previously (21). BMDMs were treated with LPS Republica Italiana no.40, February 18, 1992; National Institutes (E. coli serotype O55:B5, 100 ng/mL; Sigma-Aldrich) and mouse of Health Publication no.85-23, 1985), under the approval of the recombinant IFNg (20 ng/mL; Sigma-Aldrich) or with mouse University of L'Aquila Internal Committee and the Italian Min- recombinant IL4 (20 ng/mL; Peprotech) and mouse recombinant istry of Health. IL13 (20 ng/mL; Peprotech). The p38 inhibitors used were Vx745 Further information on the procedures used for the generation (20 mmol/L; Selleckchem), Skepinone-L (10 mmol/L; Selleck- of Gadd45bF/F mice can be found in the Supplementary chem), and SB203580 (20 mmol/L; Cell Signaling Technology). Information. For TAM isolation, intraperitoneal cells were harvested from ID8-Luc tumor-bearing mice by peritoneal lavage with 1 PBS Bone marrow chimeras þ þ þ and passed through a 70-mm cell strainer (BD Biosciences), Congenic CD45.1/Ly5.1 Gadd45b / and Gadd45b / mice following red blood cell removal with RBC lysis buffer (Sigma- were generated by crossing the corresponding C57BL/6J lines þ Aldrich). TAMs were separated using MACS MicroBead Technol- (CD45.2/Ly5.2 ) with C57BL/6J B6.SJL-Ptprca Pepcb/BoyJ mice ogy (Miltenyi Biotec) by negative selection with anti–Ly-6G (The Jackson Laboratory), carrying homozygous Ptprca (CD45.1/ MicroBead Kit, mouse, followed by positive selection with CD11b Ly5.1) alleles and used as donors in bone marrow adoptive MicroBeads, human and mouse (Miltenyi Biotec), according to transfer experiments to distinguish between hematopoietic- þ the manufacturer's instructions. derived cells of recipient (CD45.2/Ly5.2 ) and donor þ Further information on the procedures for BMDM isolation can (CD45.1/Ly5.1 ) origin. For these experiments, bone marrow be found in the Supplementary Information. cells were harvested from femurs and tibiae of 6- to 10-week old þ þ þ CD45.1/Ly5.1 Gadd45b / or Gadd45b / males under aseptic BMDM/tumor cell coculture conditions. Red blood cells were lysed using RBC lysis buffer þ þ Gadd45b / and Gadd45b / BMDMs were prepared as above (Sigma-Aldrich), washed in serum-free medium, counted, and re- and cocultured with ID8-Luc or MCA-203 tumor cells, without suspended in 1 PBS. 4 (Protocol A) or 16 (Protocol B)-week-old þ þ direct cell-to-cell contact, according to established protocols C57BL/6J Gadd45b / and Gadd45b / male recipients, carrying (18, 31). Further information on the procedures for BMDM/ Ptprcb (CD45.2/Ly5.2) alleles, were lethally irradiated with a single tumor cell coculture can be found in the Supplementary dose of 9.5 Gy and, 24 hours later, intravenously injected via the Information. tail vein with 5 106 bone marrow donor cells, to generate all four

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1277

Verzella et al.

Figure 1. The widespread correlation between elevated GADD45B expression and poor clinical outcome across human cancer types. A–M, RFS and OS in patients with the indicated malignant pathologies, representing 13 of the top 15 solid cancers for mortality worldwide and deposited in the following publicly available datasets: The Cancer Genome Atlas program (A, B, C, F, G, H, K, L,andM); the French National Cartes d'Identite des Tumeurs program (D); the Tumour Banks in the United Kingdom and Canada (E); the Chungbuk National University Hospital (I); and the Australian Ovarian Cancer Study, Royal Brisbane Hospital, Westmead Hospital and Netherlands Cancer Institute (J). Patients in each series were stratiﬁed at diagnosis in two groups on the basis of the GADD45B mRNA expression in the tumor tissues, as shown. P values are indicated.

possible chimerism combinations. Bone marrow reconstitution DEN treatment and HCC induction þ with CD45.1/Ly5.1 donor cells was veriﬁed after adoptive trans- For the initiation-only model of HCC, 15- to 17-day old þ þ fer by FACS analysis of blood leukocytes. Gadd45b / and Gadd45b / males on a C57Bl/6Jx129/SvJ or

1278 Cancer Res; 78(5) March 1, 2018 Cancer Research

Breaking Cancer Immunosuppression via GADD45b Deletion

C57Bl/6J background, as indicated, were injected intraperitone- with serum-free medium, and resuspended in 1 PBS imme- ally with a single dose of diethylnitrosamine (DEN; 5 and diately before injection. Once tumors became palpable, tumor 20 mg/kg, respectively; Sigma-Aldrich). At the indicated end- growth was monitored every other day using a vernier caliper. points, mice were sacrificed and their livers removed, photo- Tumor volumes were calculated using the formula: volume ¼ graphed, and separated into individual lobes. Externally visible A B2/2 (where A is the larger diameter, and B is the smaller tumors (0.5 mm) were measured using a caliper and counted. diameter of the tumor; ref. 24). Nineteen days after tumor cell Tumor areas were calculated using the formula: p(diameter 1 þ injection, mice were sacrificed, and tumors were removed and diameter 2)2/4 (34). For histologic evaluations, liver tissue embedded in Tissue-Tek OCT compound (Sakura Finetek) for was fixed in 10% neutral-buffered formalin solution and frozen block preparation. To investigate the macrophage-spe- paraffin embedded. The remaining liver tissue was microdissected cific role of GADD45b in sarcomagenesis, 6- to 8-week-old into tumor and nontumor tissue and stored at 80C for molec- Gadd45bF/F and Gadd45bD/D males were injected with 1.0 105 ular evaluation. exponentially growing MCA-203 cells in 200 mLofsterile1 For HCC induction in the tumor initiation–promotion model, PBS containing Matrigel (1:1; Corning). Tumor growth was mice were treated according to two distinct experimental setups, monitored using a vernier caliper, and tumor volumes were designated as Protocol A and Protocol B. As part of Protocol A, calculated as above. At the indicated endpoint, mice were þ 4-week-old C57BL/6J males (CD45.2/Ly5.2 ) were adoptively sacrificed, and tumors were photographed. þ þ transferred with bone marrow from CD45.1/Ly5.1 males, and, For the in vivo depletion of CD8 T cells, 8- to 12-week-old þ þ 4 weeks later, intravenously treated twice with of 200 mLof C57BL/6J Gadd45b / and Gadd45b / female mice were clodronate liposomes to deplete endogenous Kupffer cells. Clo- treated intraperitoneally with 250 mg of anti-CD8b antibody dronate liposomes, containing approximately 5 mg/mL of (Lyt 3.2; clone 53-5.8; Bio X Cell) or rat IgG isotype-matched clodronate, were prepared as described previously and were control antibody (clone HRPN; Bio X Cell) at day 1andday provided by N. Van Rooijen (ClodronateLiposomes.org; ref. 35). 0beforetheinjectionof1.0 105 exponentially growing Ten weeks after clodronate liposome administration, bone mar- MCA-203 cells prepared as described above, and then twice a row chimeras were intraperitoneally injected with a single dose of week with the same dose of antibody until the experimental DEN (100 mg/kg; Sigma-Aldrich) and, 4 weeks later, placed on endpoint. Tumor growth was monitored as described above. þ treatment with 0.07% of the tumor promoter, phenobarbital The efficiency of the antibody-mediated CD8 T-cell (Sigma-Aldrich), in the drinking water until the experimental depletion was determined by FACS analysis of spleen cells endpoint, according to an established tumor initiation–promo- using anti–CD8b-APC (130-106-315), anti–CD4-APC (130- tion protocol (33). As part of Protocol B, 4-week-old C57BL/6J 109-415), and anti–CD3-FITC (130-108-836) antibodies þ males (CD45.2/Ly5.2 ) were intraperitoneally injected with (Miltenyi Biotec). 75 mg/kg of DEN (Sigma-Aldrich) and, 4 weeks later, placed on For ovarian carcinoma allografts, 11- to 12-week-old C57BL/6J þ þ treatment with 0.07% phenobarbital (Sigma-Aldrich) in the Gadd45b / and Gadd45b / females were intraperitoneally drinking water, as above, to induce HCC initiation and early injected with exponentially growing 1.0 106 ID8-Luc cells in promotion. After 8 weeks, mice were adoptively transferred with 300 mL of sterile saline solution, using a 1-mL Leur-Lok Tip syringe þ bone marrow from congenic CD45.1/Ly5.1 males and, 4 weeks (BD Biosciences) with a 25-guage needle. Tumor growth was later, treated with 200 mL of clodronate liposomes, as above, to monitored weekly starting 6 weeks after tumor cell injection by deplete Kupffer cells. At the indicated endpoints, mice were bioluminescence imaging. sacrificed, and their livers removed, photographed, and separated The analyses of the macrophage-specific role of GADD45b in into individual lobes. Externally visible tumors (0.5 mm) were ovarian oncogenesis were performed essentially as described þ þ counted. For histologic evaluations, liver tissue was fixed and previously (18). Briefly, 8-week-old C57BL/6J Gadd45b / paraffin-embedded or embedded in Tissue-Tek OCT compound females (Harlan Laboratories) were intraperitoneally injected (Sakura Finetek) for frozen block preparation. Bone marrow–cell with 2 106 exponentially growing ID8-Luc cells in 300 mLof reconstitution and liver repopulation with donor-derived sterile saline solution, and tumors were allowed to grow, as above. þ CD45.1/Ly5.1 Kupffer cells were confirmed in each group of After 5 weeks, mice were randomized into two groups bone marrow chimaeras by FACS analysis of spleen and bone and intraperitoneally injected with 10 106 BMDMs from þ þ marrow cells and immunofluorescence staining of frozen liver C57BL/6J Gadd45b / or Gadd45b / mice in 300 mL of sterile sections, respectively. saline solution. ID8-Luc tumor growth was monitored weekly To validate the procedure for Kupffer cell depletion, a separate by bioluminescence imaging. Further information on the group of C57BL/6J mice was intravenously injected with a single procedures used for bioimaging can be found in the Supplemen- dose of 200 mL of clodronate or PBS liposomes. At the indicated tary Information. time points, mice were sacrificed, and their livers and spleens were removed, fixed and paraffin-embedded for immunohistologic Histopathologic analyses, IHC, immunofluorescence, and evaluation of macrophage depletion. TUNEL assays Detailed information on the procedures used for histopatho- Tumor allografts logic analyses, IHC, immunofluorescence, and TUNEL assays can For fibrosarcoma allografts, 6- to 8-week-old C57BL/6J be found in the Supplementary Information. þ þ Gadd45b / and Gadd45b / males were subcutaneously injected in the right flank with 1.0 105 MCA-203 cells in qRT-PCR analysis, Western blots, and FACS analysis 100 mLofsterile1 PBS,usinga1-mLLeur-LokTipsyringe(BD Detailed information on the procedures used for RNA extrac- Biosciences) with a 25-guage needle. MCA-203 cells were tion, qRT-PCR reactions, Western blotting, and FACS analysis can harvested from exponentially growing cultures, washed once be found in the Supplementary Information.

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1279

Verzella et al.

1280 Cancer Res; 78(5) March 1, 2018 Cancer Research

Breaking Cancer Immunosuppression via GADD45b Deletion

Statistical analysis experimental model of HCC, hepatocyte-specific Ikkb ablation Data were analyzed using GraphPad Prism version 7.0 was previously shown to promote, rather than inhibit, onco- (GraphPad Software) or R (www.R-project.org). Statistical anal- genesis (33). Therefore, to clarify the basis for the observed yses of the results were performed using either two-tailed t tests or correlation between expression of the IKKb/NF-kB–regulated two-tailed Mann–Whitney U test, depending on the distribution effector, GADD45B, and aggressive disease pathology in HCC of the data. Assumptions concerning the data (e.g., normal datasets (Fig. 1C), we sought to investigate whether GADD45b distribution and similar variation between experimental groups) was involved in this process. Fifteen-day-old Gadd45b / and þ þ were examined for appropriateness before statistical tests Gadd45b / C57BL76J males were treated with a single intra- were conducted. P values <0.05 were considered statistically peritoneal injection of the carcinogen, DEN, according to an significant. For ovarian carcinoma allografts, outliers were established protocol for HCC induction (33). As expected, all þ þ detected and removed according to the default options of the Gadd45b / mice given DEN developed typical HCCs within 9 software used. months (Fig. 2A–E). Interestingly, Gadd45b deletion signifi- Survival analyses were performed in R, using package survival. cantly reduced the number of HCCs (Fig. 2A and D). Tumor Differences between the survival distributions within each dataset surface area and maximum tumor diameter were also markedly were assessed for statistical significance using Kaplan–Meier reduced in Gadd45b / mice relative to controls (Fig. 2B and C). curves and the log-rank test, with P < 0.05 taken as the level Histologic analysis confirmed the more than 3-fold lower þ þ of significance. number of HCCs in Gadd45b / than Gadd45b / livers For the sample size of the animals used in this study, it was fixed (Fig. 2F). As typically seen in C57BL76J mice, DEN-induced in a prospective manner; no statistical method was used to tumors predominantly consisted of adenomas, with only few predetermine sample size. Mice with the indicated genotypes sporadic carcinomas, hereafter collectively referred to as HCCs. were included in the study without any randomization. Histologic Similar results were obtained using 129/SvJxC57BL/6J mice (Sup- þ þ analyses were performed in a single-blinded fashion. No blinding plementary Fig. S1A–S1F). Notably, Gadd45b / and Gadd45b / was used for the remaining analyses. HCCs displayed no difference with respect to tumor grade, mitotic index, or necrosis grade, and similar percentages of proliferating Results and apoptotic cells (Supplementary Fig. S1G–S1J). Likewise, the GADD45B analysis of cell-cycle regulators and oncogenic factors showed no expression denotes aggressive malignant disease / þ/þ across most human cancers difference between Gadd45b and Gadd45b livers when We recently identified the product of the NF-kB–regulated comparing corresponding tumor or nontumor tissues (Supple- fi gene, GADD45B, as an essential survival factor and novel thera- mentary Fig. S1K). Hence, opposite to hepatocyte-speci c Ikkb peutic target in multiple myeloma (24). We therefore investigated loss (33), Gadd45b ablation inhibits DEN-induced hepatocellular whether GADD45B was involved in any types of malignancy carcinogenesis, while having no effect on tumor cell proliferation beyond multiple myeloma. Strikingly, in publicly available or apoptosis. patient datasets, elevated GADD45B expression correlated with rapid disease progression in 13 of the top 15 solid cancers for Gadd45b loss augments tumor-associated macrophage and mortality worldwide (36). When patients with prevalent disease T-cell infiltration and TLS formation subtypes were stratified at diagnosis on the basis of GADD45B Despite the otherwise similar histologic and molecular char- expression in the tumor tissues, the patient cohorts expressing acteristics (Fig. 2E; Supplementary Figs. S1E and S1G–S1K), / high GADD45B levels exhibited significantly shorter RFS and/or Gadd45b HCCs contained considerably more and larger þ/þ OS than the corresponding cohorts expressing low GADD45B immunoinflammatory infiltrates than Gadd45b tumors mRNA levels (Fig. 1A–M). Collectively, these results identify (Fig. 2G; see also Fig. 2E and H; Supplementary Fig. S1E). / þ/þ GADD45b as a hallmark of aggressive pathology across the most Strikingly, in Gadd45b , but not Gadd45b HCCs, intratu- prevalent human cancer types and postulate its general clinical moral infiltrates were also often organized into follicle-like aggre- significance in malignant disease. gates, reminiscent of tertiary lymphoid structures (TLS; Fig. 2G–J; see also Fig. 2E; Supplementary Figs. S1E and S2A and S2B), Gadd45b loss reduces chemically induced hepatocellular denoting sites of thriving adaptive immune reactions and favor- carcinogenesis able clinical outcome in multiple human cancers (37). Consis- þ þ TheroleofIKKb/NF-kB in hepatocellular carcinogenesis tently, whereas only 33% of Gadd45b / mice contained at least is controversial (16, 17). Indeed, in the widely used DEN one HCC featuring TLS-like aggregates (hereafter referred to

Figure 2. Reduced DEN-induced HCC development with increased intratumoral immunoinflammatory infiltrates and TLS formation in Gadd45b/ mice. A–C, Number of tumors (0.5 mm; A), tumor surface area (B), and maximum tumor diameter (C) in livers of Gadd45bþ/þ (n ¼ 9) and Gadd45b/ (n ¼ 12) C57BL/6J males 9 months after DEN injection (20 mg/kg). Values, means SEM. D, Gross liver morphology in representative mice from A–C. Arrowheads, tumors. E, Images of hematoxylin and eosin (H&E) staining showing the liver histology in representative mice from A–C. Hatched lines (top), tumor areas. Solid lines (top), areas magnified in the bottom. Arrowhead (bottom, right), a typical immunoinflammatory aggregate. Scales and magnifications are shown. F, Number of histologically confirmed HCCs per examined liver sections from individual mice in A–C. Each symbol represents an individual mouse. Horizontal lines, means. G and H, Percentages of HCCs (G) and combined HCCs and preneoplastic foci (H) containing TLSs in Gadd45bþ/þ (n ¼ 19) and Gadd45b/ (n ¼ 26) 129/SvJxC57BL/6J males 11 months after DEN injection (5 mg/kg). Each symbol represents an individual mouse. Horizontal lines, means. I, IHC analysis showing the percentage of positive area or number of positive cells per field at 400, as stated, of the indicated immune cell populations in HCCs from G. Values, means SEM (IBA-1: Gadd45bþ/þ,n¼ 47; Gadd45b/,n¼ 34. CD8: Gadd45bþ/þ,n¼ 37; Gadd45b/,n¼ 33. All other markers: Gadd45bþ/þ,24 n 26; Gadd45b/,n¼ 26 or n ¼ 27). J, Images of hematoxylin and eosin and IHC staining of representative tumor sections from I. Scales and magnifications are shown. A–C and F–I, , P < 0.05; , P < 0.01; , P < 0.001. See also Supplementary Figs. S1 and S2.

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1281

Verzella et al.

as TLSs), well-organized TLSs were often found in multiple HCCs therefore examined the effect of Gadd45b loss on the acute liver þ þ in 80% of Gadd45b / mice (Supplementary Fig. S2C; see also response to DEN administration. As expected, in Gadd45b / Supplementary Fig. S2D). mice, DEN injection induced hepatocyte apoptosis at 48 hours þ þ In both Gadd45b / and Gadd45b / HCCs, intratumoral leu- (Supplementary Fig. S3A and S3B). The number of apoptotic kocyte infiltrates predominantly consisted of lymphocytes and hepatocytes was modestly increased, in Gadd45b / mice after macrophages, with only few sparse granulocytes (Fig. 2I and J). DEN administration, but not at baseline, compared with controls However, Gadd45b / HCCs contained considerably more (Supplementary Fig. S3A and S3B). Circulating liver enzymes, þ þ F4/80 and IBA-1 macrophages and B and T lymphocytes than indicative of hepatocyte damage, were also higher in DEN-treated þ þ þ þ Gadd45b / HCCs (Fig. 2I and J; see also Fig. 3A and B). Crucially, Gadd45b / than Gadd45b / mice (Supplementary Fig. S3C), þ Gadd45b loss also increased the intratumoral infiltration by CD8 even though the livers of these mice displayed similarly low T cells, a hallmark of effective antitumor immune responses numbers of dead cells (Supplementary Fig. S3D and S3E). (Fig. 2I and J). Interestingly, whereas the increased numbers of Because DEN-induced hepatocyte toxicity involves JNK activa- þ IBA-1 macrophages and lymphocytes observed in Gadd45b / tion (33, 39), and GADD45b has the capacity to suppress JNK HCCs were mostly due to the greater frequency and larger size of signaling in the liver (32), we investigated the effects of Gadd45b þ TLSs, the increased intratumoral F4/80 macrophage numbers loss on MAPK activation by DEN. As shown in Supplementary Fig. þ þ were mainly attributable to the higher density of these cells S3F, DEN induced strong JNK activation in Gadd45b / livers, but outside TLSs. Plausibly, this distinct effect of Gadd45b loss on this activation was unaffected by Gadd45b loss. Gadd45b deficien- þ the F4/80 macrophage distribution within the TME reflects the cy also had no effect on DEN-induced ERK, p38 or STAT3 central role of these cells in sensing "danger" signals and antigens signaling, nor did it have any effect on hepatocyte proliferation, arising from tumor cells and processing them into cues to elicit either at baseline or following DEN injection (Supplementary reciprocal adaptive immune reactions (38). Irrespective of the Fig. S3G and S3H). Hence, Gadd45b loss mildly augments DEN- mechanism(s), our data demonstrate that Gadd45b loss pro- induced liver damage, but does not affect hepatocyte proliferation foundly affects tumor-associated inflammation and macrophage and is, therefore, unlikely to impact HCC development by affect- and lymphocyte infiltration within DEN-induced HCCs. ing the acute liver response to DEN. Because resident liver macrophages, so-called Kupffer cells, play a central role in DEN-induced carcinogenesis through a Hematopoietic cell–specific Gadd45b loss reduces DEN- mechanism that depends on IKK/NF-kB (17, 33), we further induced carcinogenesis investigated the effect of GADD45b on TAMs. In keeping with Given the association between diminished HCC burden and their heightened state of inflammation, Gadd45b / HCCs con- increased intratumoral immunoinflammatory infiltration tained a significantly higher number of myeloid-like cells expres- observed in Gadd45b / mice, we hypothesized that GADD45b sing inflammatory markers, such as inducible nitric oxide promotes hepatocellular carcinogenesis by operating within synthase (iNOS) and cyclo-oxygenase-2 (COX-2), than immune and inflammatory cells. To distinguish between the þ þ Gadd45b / HCCs (Fig. 3A and B; Supplementary Fig. S2E; oncogenic roles of GADD45b in the TME and the hepatic paren- þ ref. 14). They also contained markedly more IBA-1 macrophages chyma, we generated bone marrow chimeras harboring all pos- within well-organized TLSs expressing MHC-II, a key antigen- sible combinations of tissue-specific Gadd45b deficiency in the presenting molecule and prototypical proinflammatory activa- parenchymal tissue and bone marrow–derived cells (Fig. 4A). tion marker (Fig. 3C and D; see also Fig. 3A and B; Supplementary Given the importance of Kupffer cells in driving DEN-induced Fig. S2E; ref. 7), thus phenocopying the effects of myeloid-asso- carcinogenesis (17, 33), chimeric mice were then treated with ciated IKKb/NF-kB inhibition on TAMs (18, 19). Underscoring clodronate liposomes to deplete endogenous Kupffer cells and the significance of these findings, Gadd45b / HCCs additionally thereby permit liver repopulation with donor-derived Kupffer displayed considerably more TLS-associated cells expressing the cells, following which, mice were administered DEN, along with immunoinhibitory enzyme, indoleamine-2,3-dioxygenase the tumor promoter, phenobarbital (Protocol A; Fig. 4A), accord- (IDO; Figs. 3E and F), suggesting that the upregulation of immune ing to an established tumor initiation–promotion protocol (33). checkpoint molecules, such as IDO (3, 4), could conceal the full Successful bone marrow–cell reconstitution and host Kupffer immunostimulatory potential of Gadd45b loss. In contrast, cell liver repopulation with donor-derived Kupffer cells were Gadd45b deletion had no effect on alternative TAM activation, confirmed by FACS and immunofluorescence analyses, respec- which is more typical of established neoplasias (Figs. 3A and B; tively (Supplementary Fig. S4A-S4B; see also Supplementary see also Supplementary Fig. S2E; refs. 7, 8). Hence, Gadd45b loss Fig. S4C). As expected, DEN-treated chimeras bearing isogenic þ/þ / enhances proinflammatory TAM activation within HCCs. Collec- Gadd45b or Gadd45b tissues (WTBM:WTHost and KOBM: tively, these findings indicate that GADD45b plays an essential KOHost, respectively) developed similar numbers of HCCs as their role in curbing proinflammatory TAM activation, intratumoral nontransplanted counterparts (Figs. 4B and C; see also Fig. 2A; lymphocyte infiltration and TLS formation within established Supplementary Fig. S1A). Notably, however, the ratio between neoplasias. HCC numbers in these mice was more than 5-folds (Fig. 4B; compare WTBM:WTHost with KOBM:KOHost) relative to the approx- Gadd45b loss augments DEN-induced hepatocyte toxicity, but imately 2-fold ratio observed in mice treated according to not hepatocyte proliferation the tumor initiation–only protocol (Fig. 2A; Supplementary Although in established tumors macrophage-driven inflamma- Fig. S1A), suggesting that Gadd45b loss has a greater impact on tion has the potential to elicit antitumor immune responses, it HCC progression than on tumor initiation (discussed below). was shown that when it ensues in response to acute DEN-induced Strikingly, selective Gadd45b ablation in bone marrow–derived liver damage, this inflammation triggers compensatory hepato- cells reduced the numbers of HCCs to similar levels as in isogenic / cyte proliferation, thereby facilitating HCC initiation (33). We Gadd45b mice (Fig. 4B; compare KOBM:WTHost with WTBM:

1282 Cancer Res; 78(5) March 1, 2018 Cancer Research

Breaking Cancer Immunosuppression via GADD45b Deletion

Figure 3. Gadd45b loss increases macrophage infiltration and proinflammatory TAM activation within HCCs. A, IHC analysis showing the percentage of positive area or number of positive cells per field at 400, as stated, for the indicated macrophage (F4/80, IBA-1) and proinflammatory and anti-inflammatory activation markers in HCCs from Fig. 2A–C. B, Images of IHC staining of representative tumor sections from A. Images of Gadd45b/ HCCs are shown for nonorganized immunoinflammatory cell clusters (middle) and well-organized TLSs (right). Insets denote magnified areas at the top right corners. C, Immunofluorescence analysis showing the percentages of IBA-1þ macrophages expressing MHC-II per field at 400 in HCCs from Fig. 2A–C. D, Images of immunofluorescence staining of representative tumor sections from C. Red, anti–IBA-1; green, anti–MHC-II; blue, DAPI. E, IHC analysis showing the percentage of IDO-positive area per field at 400 in Gadd45bþ/þ and Gadd45b/ HCCs from A. F, Images of IHC staining of representative tumor sections from E. A, C,andE, Values denote means SEM (Gadd45bþ/þ,14 n 30; Gadd45b/,12 n 15). , P < 0.05; , P < 0.01; , P < 0.001. B, D, and F, Scales and magnifications are shown. See also Supplementary Figs. S2 and S3.

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1283

Verzella et al.

Figure 4. Reduction of DEN-induced HCC development by Gadd45b deletion in bone marrow–derived cells. A, Summary of the treatment schedule used for Protocol A.

B, Numbers of tumors (0.5 mm) in livers of WTBM:WTHost (n ¼ 40), KOBM:KOHost (n ¼ 25), KOBM:WTHost (n ¼ 36), and WTBM:KOHost (n ¼ 35) bone marrow chimeras at the time shown in A. C, Gross liver morphology in representative mice from B. D, Summary of the treatment schedule used for Protocol B.

E, Numbers of tumors (0.5 mm) in livers of WTBM:WTHost (n ¼ 24), KOBM:KOHost (n ¼ 23), KOBM:WTHost (n ¼ 24), and WTBM:KOHost (n ¼ 29) bone marrow chimeras at the time shown in D. F, Gross liver morphology in representative mice from E. A–F, WT, Gadd45bþ/þ; KO, Gadd45b/. Bone marrow chimeras are identiﬁed as per "bone marrow donor:recipient" genotypes. B and E, Values, means SEM. , P < 0.05; , P < 0.01; , P < 0.001; , P < 0.0001. C and F, Arrowheads, tumors. Also, see Supplementary Fig. S4.

WTHost; also compare KOBM:WTHost with KOBM:KOHost). In con- Gadd45b loss hinders HCC progression trast, selective Gadd45b deletion in the hepatic parenchyma did Given that the effect of Gadd45b loss on DEN-induced carci- not significantly affect HCC burden (compare WTBM:KOHost with nogenesis intensifies in a tumor initiation–promotion model WTBM:WTHost). In keeping with these findings, the additional (Fig. 4B and C), is associated with increased TME-based inflam- Gadd45b ablation in bone marrow–derived cells did result in a mation (Figs. 2G–J and 3A–F; Supplementary Fig. S2A–S2E), and significant reduction in tumor numbers in Gadd45b / hosts does not affect the acute hepatocyte proliferative response to DEN (compare WTBM:KOHost with KOBM:KOHost). Hence, selective (Supplementary Fig. S3G and S3H), we hypothesized that Gadd45b loss in hematopoietic-derived cells is sufficient on its GADD45b contributes to HCC progression, rather than initiation. own to fully recapitulate the phenotype of Gadd45b / mice in As part of so-called "Protocol B" (Fig. 4D), mice of both genotypes DEN-induced carcinogenesis. We concluded that GADD45b pro- were first challenged with DEN and placed on treatment with motes this oncogenic process largely by operating within immune phenobarbital, and then adoptively transferred with Gadd45b / þ þ and inflammatory cells. or Gadd45b / bone marrow, followed by Kupffer cell depletion,

1284 Cancer Res; 78(5) March 1, 2018 Cancer Research

Breaking Cancer Immunosuppression via GADD45b Deletion

as for Protocol A. Because, the Gadd45b status of inflammatory rapid evolution of fibrosarcoma allografts relative to endogenous cells is modified 3 months after DEN injection, when DEN-induced HCCs and the delayed onset of adaptive relative to preneoplastic foci have already largely progressed to HCCs innate immune reactions (see also Figs. 2I and J). Indeed, the þ (40), we reasoned that Protocol B would allow to distinguish depletion of cytotoxic CD8 T cells by the use of a CD8-specific between roles of GADD45b in tumor progression versus roles in antibody resulted in a marked increase in fibrosarcoma growth, tumor initiation. completely ablating any inhibitory effect of Gadd45b loss on Successful bone marrow–cell engraftment and Kupffer cell liver oncogenesis (Fig. 5C and D; Supplementary Fig. S5B; compare repopulation with donor-derived cells were verified as for Proto- tumor growth in Gadd45b / mice treated with anti-CD8 vs. col A (Supplementary Fig. S4D and S4E). As expected, 12 months control antibody; also note the similar tumor growth in þ þ after DEN injection, the ratio between HCC numbers of isogenic Gadd45b / and Gadd45b / mice treated with the anti-CD8 þ þ Gadd45b / and Gadd45b / mice was similar to that observed antibody; see also Supplementary Fig. S5C). These data unequiv- þ with Protocol A (Fig. 4E and F; compare KOBM:KOHost with WTBM: ocally demonstrate the critical importance of cytotoxic CD8 WTHost; see also Fig. 4B and C). Strikingly, even when introduced T cells in the antitumor effect of TME-associated Gadd45b loss. þ þ after tumor initiation had already largely occurred in Gadd45b / In contrast, TME-infiltrating B cells and granulocytes were unaf- livers, bone marrow–derived cell-specific Gadd45b ablation fected by Gadd45b loss. retained a virtually intact capacity to diminish oncogenesis (Fig. Underscoring the significance of the effects of GADD45b on 4E; compare KOBM:WTHost with WTBM:WTHost). Accordingly, TAMs, these cells comprised the major cellular component of the concomitant Gadd45b ablation in the hepatic parenchyma had immunoinflammatory infiltrate in the Gadd45b / TME. Strik- / no additional effect on HCC burden (compare KOBM:WTHost with ingly, the Gadd45b TME also contained an approximately þ/þ KOBM:KOHost). Reciprocally, the introduction of Gadd45b 4-fold higher number of TAMs expressing the proinflammatory þ þ bone marrow into Gadd45b / hosts, 3 months after DEN injec- markers, iNOS and COX-2 (14), than the Gadd45b / TME tion, was still effective in restoring oncogenesis (compare WTBM: (Figs. 5E; Supplementary Fig. S5D). Reciprocally, fibrosarcomas / KOHost with KOBM:KOHost). Notably, these results were broadly from Gadd45b mice displayed a lower number of Ym-1/ þ þ þ equivalent to those obtained with Protocol A (see Fig. 4B). Chi3l3 TAMs than tumors from Gadd45b / mice. Other anti- Collectively, these findings reinforce the conclusions that the inflammatory activation markers were instead unaffected by GADD45b oncogenic function in DEN-induced tumorigenesis is Gadd45b loss (14). Hence, TME-associated Gadd45b ablation largely restricted to the hematopoietic system and impacts augments TAM infiltration, proinflammatory TAM activation, and þ upon tumor progression, while likely having little impact upon CD8 T-cell–driven antitumor immune responses in fibrosarco- tumor initiation. ma, thus reinforcing our findings with DEN-induced HCC.

TME-associated Gadd45b loss in fibrosarcoma augments TAM TME-specific Gadd45b loss augments proinflammatory TAM infiltration and proinflammatory activation, while reducing activation and intratumoral T-cell infiltration, while reducing tumor growth ovarian oncogenesis To clarify the basis for the tumor-promoting activity of To further understand the general significance of GADD45b in GADD45b in the TME and its significance in the association TAM activation and malignant progression, we examined a third between GADD45B expression and aggressive disease pathology cancer model in which IKK/NF-kB was shown to drive oncogen- across human cancer types (Figs. 1A–M), we sought to investigate esis by operating in myelomonocytic cells. We used ID8-Luc other model systems of established neoplasia, in which IKK/ ovarian adenocarcinoma allografts because previous studies had NF-kB drives oncogenesis by operating in immunoinflammatory shown that macrophage-specific IKKb inhibition in this model cells. We reasoned that this approach could additionally clarify reverses the typical anti-inflammatory TAM phenotype and the basis for the oncogenic function of IKK/NF-kB itself in these enhances proinflammatory TAM activation, thereby causing cells. Indeed, although IKK/NF-kB inhibition in different epithe- tumor regression (18). Consistently, whereas syngeneic ID8-Luc þ þ lial cell types was shown to have discordant outcomes on onco- tumors grew rapidly in the peritoneum of Gadd45b / mice, genesis, this inhibition in inflammatory cells, especially those of tumor growth was markedly reduced in Gadd45b / mice (Figs. the myeloid lineage, has invariably resulted in oncogenesis sup- 5F and G), thus demonstrating that TME-specific Gadd45b abla- pression (16). We chose an initiated fibrosarcoma model in which tion also curbs ovarian oncogenesis. hematopoietic cell–specific NF-kB1/p105 ablation was shown to Notably, intratumoral leukocyte infiltrates yielded a higher augment TME-based inflammation and proinflammatory TAM number of TAMs, T cells and granulocytes, but not B cells, in þ þ activation, resulting in diminished tumor growth (19). As Gadd45b / compared with Gadd45b / mice (Fig. 5H). Signif- expected, subcutaneously implanted MCA-203 fibrosarcomas icantly, as seen in DEN-induced HCC (Figs. 2I and J), the relative þ þ grew rapidly in immunocompetent C57BL76J Gadd45b / mice T-cell abundance in the Gadd45b / TME was largely due to an þ (Fig. 5A). Notably, however, tumor growth was markedly reduced increased influx of CD8 cytotoxic T cells, which drive effective in Gadd45b / hosts, thereby confirming in a different model antitumor immune reactions (Fig. 5H; refs. 3, 5), rather than þ system the essential oncogenic role of GADD45b in the TME. CD4 T cells. Consistent with the increased immunoreactivity of As seen with HCC (Figs. 2I and J and 3A–D), the Gadd45b / the Gadd45b / TME, a larger proportion of Gadd45b / than þ þ þ þ TME contained a higher number of TAMs than the Gadd45b / Gadd45b / TAMs expressed high surface levels of MHC-II (Fig. 5I TME (Fig. 5B; Supplementary Fig. S5A). Fibrosarcomas from and J; Supplementary Fig. S5E). Purified Gadd45b / TAMs also Gadd45b / mice also contained more T cells than tumors from expressed higher levels of other proinflammatory genes, including þ þ þ þ Gadd45b / hosts, although at the early time point examined (i.e., Nos2, Cxcl9, Cxcl10, Tnfa, and Il6, than Gadd45b / TAMs, d19), this T-cell increase had not reached statistical significance whereas anti-inflammatory activation genes, with the exception (Figs. 5B; Supplementary Fig. S5A). Plausibly, this reflected the of Ym-1/Chi3l3, were unaffected by Gadd45b loss (Fig. 5K;

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1285

Verzella et al.

1286 Cancer Res; 78(5) March 1, 2018 Cancer Research

Breaking Cancer Immunosuppression via GADD45b Deletion

Supplementary Fig. S5F; ref. 14). In keeping with their heightened further significantly upregulated in IFNg/LPS-treated Gadd45b / þ þ state of immune activation, Gadd45b / TAMs additionally BMDMs compared with Gadd45b / BMDMs. Immune check- expressed higher levels of immune checkpoint molecule- point molecule-coding genes, including Ido and Pd-l2, were also encoding genes, for example, Ido, Ctla-4, Pd-l1, and B7-h3 markedly increased in IFNg/LPS-stimulated Gadd45b / BMDMs (Fig. 5L; refs. 3, 4), thus confirming the observations made with relative to controls (Fig. 6B), in keeping with our in vivo findings DEN-induced HCC (Fig. 3E and F). Hence, in three distinct (Figs. 3E and F and 5L). In contrast, Gadd45b expression was models of solid cancer, TME-specific Gadd45b ablation increased unaffected by treatment with IL4 and IL13, as was the upregula- proinflammatory TAM activation, and intratumoral macrophage tion of each of the anti-inflammatory M2-like genes analyzed þ and CD8 T-cell infiltration (see also Figs. 2G–J, 3A–F, 5B, E, and (Fig. 6C). We concluded that GADD45b is required to curb H–L; Supplementary Figs. S2E; S5A, S5D, and S5E), resulting in inflammation and M1-like macrophage activation, but is not diminished tumor growth. Interestingly, these effects of Gadd45b directly involved in anti-inflammatory M2-like polarization, sug- loss on oncogenesis phenocopied the effects of myeloid-specific gesting that the perturbation of the anti-inflammatory activation IKKb/NF-kB inhibition (16, 18, 19). Collectively, these findings profile of Gadd45b / TAMs, occasionally observed in coculture underpin the importance of the TME-specific role of GADD45b in systems and in vivo, stemmed from the skewing of these cells curbing tumor-associated inflammation and immune infiltration toward a proinflammatory activation phenotype. Collectively, across different cancer types. these findings provide a mechanism for the effect of Gadd45b loss on tumor-associated inflammation and proinflammatory Gadd45b loss promotes proinflammatory macrophage TAM polarization, in vivo. activation via a cell-autonomous mechanism We sought to investigate whether GADD45b mediated its Gadd45b loss exacerbates proinflammatory M1-like effects on TAM activation via a cell-autonomous mechanism. We polarization by enhancing p38 signaling used an ex vivo coculture model system mimicking the tumor cell/ To further the mechanistic understanding of the role of macrophage interactions occurring in cancer, in vivo (18). Con- GADD45b in macrophage polarization, we examined STAT1 sistent with our in vivo findings (Figs. 5E; Supplementary and MAPK signaling, which upregulates the M1-like activation Fig. S5D), upon coculture with MCA-203 cells, Gadd45b / program downstream of IFNg receptor (IFNgR) and TLRs (41). BMDMs expressed markedly higher levels of proinflammatory AsshowninFig.6DandE,stimulationwithIFNg and LPS M1-like genes, including Nos2, Tnfa, and Cxcl10, and lower levels activated these pathways, as well as GADD45b expression, in þ þ of Arg-1, but not of other anti-inflammatory M2-like genes, than Gadd45b / BMDMs. Yet, Gadd45b deletionhadnoeffecton þ þ Gadd45b / BMDMs (Fig. 6A). Similar results were obtained either STAT1 or ERK phosphorylation, nor did it have any þ þ when Gadd45b / and Gadd45b / BMDMs were cocultured effect on JNK signaling, nor on NF-kB activation (Fig. 6D and under similar conditions with ID8-Luc carcinoma cells (Supple- E). Surprisingly, Gadd45b loss instead markedly augmented mentary Fig. S6A; see also Figs. 5I–K; Supplementary Fig. S5F). and prolonged IFNg/LPS-induced p38 signaling, at each of Hence, GADD45b plays an essential and cell-autonomous role in the time points examined (Fig. 6D). We concluded that restraining proinflammatory macrophage activation in response GADD45b is required to restrict both the magnitude and to cues from tumor cells. duration of p38 activation during proinflammatory macro- To clarify the mechanisms by which GADD45b governs tumor phage polarization. cell–induced macrophage activation, we used a tumor cell–free Because p38 governs inflammation and expression of many model system, which employs cytokines and TLR ligands to of the proinflammatory M1-denoting genes deregulated by polarize macrophages (7, 8). As expected, proinflammatory Gadd45b loss (42), we hypothesized that GADD45b inhibits M1-like genes, as well as known NF-kB target genes, including proinflammatory macrophage polarization by restraining exac- Gadd45b, Nfkbia, and Tnfaip3, were strongly upregulated by IFNg erbated p38 signaling. We therefore investigated whether atten- þ þ and lipopolysaccharide (LPS) in Gadd45b / BMDMs (Fig. 6B). uating this signaling reversed the effects of Gadd45b loss on Each of the proinflammatory M1-like genes tested, however, was macrophage activation. Notably, treatment with the specific

Figure 5. Increased intratumoral immunoinflammatory infiltration, proinflammatory TAM activation, and reduced fibrosarcoma and ovarian carcinoma growth in Gadd45b/ mice. A, Volumes of subcutaneous MCA-203 fibrosarcoma allografts in Gadd45bþ/þ (n ¼ 10) and Gadd45b/ (n ¼ 9) mice at the times shown. B, IHC analysis showing the percentage of positive areas per field at 400 of the indicated immune cell populations in MCA-203 tumors from Gadd45bþ/þ (n ¼ 7) and Gadd45b/ (n ¼ 8) mice 19 days after tumor cell injection. C, Volumes of subcutaneous MCA-203 fibrosarcoma allografts in Gadd45bþ/þ and Gadd45b/ mice treated with anti-CD8 or isotype control antibody, as shown, at the times indicated. Gadd45bþ/þ: anti-CD8 (n ¼ 9), isotype (n ¼ 7); Gadd45b/: anti-CD8 (n ¼ 9), isotype (n ¼ 4). D, Images of representative tumors from C at day 20. E, IHC analysis showing the percentage of positive areas per field at 400 of the indicated proinflammatory and anti-inflammatory activation markers in the tumors from B. F, Relative luminescence units (RLU) of intraperitoneal ID8-Luc ovarian tumor allografts in Gadd45bþ/þ (n ¼ 10) and Gadd45b/ (n ¼ 10) mice at the times shown after tumor cell injection. G, Bioluminescence images of representative mice from F presented as a pseudocolor scale, whereby red and blue denote the highest and lowest photon flux, respectively. H, FACS analysis showing the percentage of the indicated intratumoral immune cell populations in Gadd45bþ/þ (n ¼ 12) and Gadd45b/ (n ¼ 13) mice 60 days after ID8-Luc tumor cell injection. I, FACS analysis showing the percentage of CD11bþ and F4/80þ double positive TAMs expressing high MHC-II levels (MHC-IIhigh) in ID8-Luc tumors from H. J, FACS analysis showing the median MHC-II fluorescence intensity of CD11bþ and F4/80þ double positive TAMs in ID8-Luc tumors from H. H–J, Boxes span between the highest values of the first and third quartiles; whiskers extend to the highest and lowest values within 1.5 of the interquartile range. Lines within boxes, medians. K and L, qRT-PCR showing the relative mRNA levels of the indicated proinflammatory activation markers (K) and immune checkpoint molecules (L) in CD11bþ TAMs from Gadd45bþ/þ (n ¼ 22) and Gadd45b/ (n ¼ 10 or n ¼ 12) mice 10 weeks after ID8-Luc cell injection. A–C, E, F, K,andL, Values, means SEM. A–C, E, F,andH–L, , P < 0.05; , P < 0.01; , P < 0.001. Also, see Supplementary Fig. S5.

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1287

Verzella et al.

Figure 6. Gadd45b loss increases proinflammatory macrophage activation by enhancing p38 signaling. A, qRT-PCR showing the relative mRNA levels of the indicated proinflammatory and anti-inflammatory genes in BMDMs from Gadd45bþ/þ and Gadd45b/ mice after a 24-hour coculture with MCA-203 cells. B, qRT-PCR showing the relative mRNA levels of the indicated proinflammatory, NF-kB–regulated and immune checkpoint molecule-coding genes in BMDMs from Gadd45b/ and Gadd45bþ/þ mice after a 4-hour (Nfkbia, Tnfaip3,andGadd45b) or 12-hour (all other genes) stimulation with LPS and IFNg. C, qRT-PCR showing the relative mRNA levels of the indicated anti-inflammatory genes after a 26-hour stimulation with IL4 and IL13. D and E, Western blots showing þ/þ / total and phosphorylated (P) proteins in Gadd45b and Gadd45b BMDMs after stimulation with LPS and IFNg for the times indicated. b-Actin is shown as loading control. F, qRT-PCR showing the relative mRNA levels of the indicated proinflammatory genes in Gadd45b/ BMDMs left untreated () or treated with LPS and IFNg (þ) for 12 hours in the presence (þ) or absence () of the p38 inhibitor, Vx745 (20 mmol/L). A–C and F, Values, means SEM [A, B (top), and F, n ¼ 4; B (bottom) and C, n ¼ 3]. , P < 0.05; , P < 0.01; , P < 0.001. Also, see Supplementary Fig. S6.

1288 Cancer Res; 78(5) March 1, 2018 Cancer Research

Breaking Cancer Immunosuppression via GADD45b Deletion

p38a/b inhibitor, Vx745 (http://www.kinase-screen.mrc.ac.uk/ Discussion screening-compounds/348786), downregulated the IFNg/LPS– We have identified an essential innate immunity "checkpoint" dependent induction of each of the proinflammatory M1-like governed by the GADD45b-dependent axis of the NF-kB pathway genes that were overactivated by Gadd45b loss to levels similar þ þ þ that counters TME-based inflammation and CD8 T-cell traffick- to those observed in IFNg/LPS–stimulated Gadd45b / BMDMs ing into tumors, the major barrier to effective anticancer immu- (Fig. 6F; see also Fig. 6B). Comparable results were obtained notherapy (3, 5). In three distinct models of cancers that are using two additional, structurally unrelated p38 inhibitors, that largely refractory to immunotherapies (43–45), myeloid-associ- is, skepinone-L and SB203580, thus excluding any off-target ated Gadd45b ablation restored proinflammatory TAM activation effect of Vx745 (Supplementary Fig. S6B and S6C). We con- þ and intratumoral CD8 T-cell infiltration, leading to diminished cluded that GADD45b suppresses proinflammatory macro- tumor growth. These findings uncover a mechanism for the phage activation by selectively downregulating IFNg/LPS– immunosuppressive activity of the TME that fosters oncogenesis. induced p38 signaling. They also provide a basis for the anti-inflammatory role that IKKb/ NF-kB plays in the myeloid lineage and identify a candidate Macrophage-specific Gadd45b loss is sufficient to diminish therapeutic route in the innate immune system for "reeducating" oncogenesis TAMs to overcome TME-mediated immunosuppression and, To clarify the cause–effect relationship between macrophage- potentially, resistance to immunotherapies. specific Gadd45b deficiency and oncogenesis suppression, we The treatment of certain malignancies is being revolutionized generated mice carrying homozygous loxP-flanked Gadd45b by the development of effective immunotherapies, such as alleles (Gadd45bF/F mice; Fig. 7A; Supplementary Fig. S7A antagonists of immune checkpoint molecules (4). Yet, the large and S7B), and crossed them with LysM-cre transgenic mice to majority of patients fail to respond to these treatments, due to the obtain myeloid-specific Gadd45b-null mice (Gadd45bDM/DM), after presence of additional immunoinhibitory mechanisms in the Cre-mediated recombination (Supplementary Fig. S7C; ref. 33). TME that prevent T-cell access into tumors or cause T-cell exhaus- PCR analyses confirmed the selective excision of the Gadd45bF tion (3, 5). However, the limited understanding of these mechan- alleles and the resulting ablation of Gadd45b mRNA expression isms currently precludes the development of effective treatment in Gadd45bDM/DM splenic macrophages, but not splenic B or strategies for breaking the TME-mediated resistance to immu- T lymphocytes (Fig. 7B and C). notherapies. An attractive approach to overcome this resistance Given that the LysM-cre transgene, which efficiently deletes would be to harness the innate immune system to induce tumor- loxP-targeted alleles in circulating myelomonocytic cells, is not þ based inflammation and CD8 T-cell entry and expansion in the useful in Kupffer cells (33), Gadd45bDM/DM mice could not be TME. Together, the myeloid-specific role of IKKb/NF-kB in sup- tested in the DEN HCC model. We therefore used them in the pressing proinflammatory TAM activation (15) and the central context of MCA-203 fibrosarcoma allografts. As expected, role of myeloid cells in cancer immune evasion (6, 7) provide a MCA-203 tumor growth was unaffected by the presence of compelling rationale for therapeutically targeting IKKb/NF-kB Gadd45bF/F alleles (Fig. 7D and E; see also Fig. 5A). Strikingly, signaling to reprogram TAMs toward a proinflammatory activa- however, this growth was markedly diminished in Gadd45bDM/DM tion phenotype capable of eliciting local antitumor immune mice to a similar extent as seen in Gadd45b / mice (Fig. 7D responses (18, 19). Yet, no specific IKKb/NF-kB inhibitor has so and E; see Fig. 5A), indicating that macrophage-specific Gadd45b far been clinically approved. loss is sufficient on its own to fully recapitulate the effect Our finding that GADD45b mediates an essential innate of complete TME-based Gadd45b deficiency on inflammation- immunoinhibitory mechanism therefore provides an attractive driven oncogenesis. þ route downstream in the IKKb/NF-kB pathway to redirect CD8 We sought to confirm this macrophage-specific role of þ T-cell trafficking into tumors and reactivate silenced CD8 T-cell– GADD45b in oncogenesis by using an alternative mouse model. driven antitumor immune responses. Congruently, in various To this end, ID8-Luc ovarian tumors were allowed to grow in þ þ cancer types, selective GADD45b ablation in myelomonocytic Gadd45b / mice until they became detectable by bioimaging, at cells closely phenocopied the effects of IKKb/NF-kB inactivation which point, mice were adoptively transferred with ex vivo cul- þ/þ / on oncogenesis (15, 16, 18, 19). Importantly, myeloid-specific tured Gadd45b or Gadd45b BMDMs (0 week; Figs. 7F–H). þ þ þ GADD45b deletion appeared to override TME-mediated CD8 T- As expected, following Gadd45b / BMDM transfer, ovarian cell exclusion, plausibly through a mechanism involving the tumors continued to grow rapidly (Figs. 7G and H). In contrast, upregulation of CXCR3-binding chemokines, for example, the transfer of Gadd45b / BMDMs resulted in a significant CXCL9, CXCL10, and CXCL11, which recruit Th-1–type and inhibition of tumor growth. As seen with fibrosarcoma (Figs. þ CD8 effector T cells into tumors and extranodal sites of inflam- 7D and E), the effects of macrophage-specific Gadd45b loss on mation (5, 46). Underscoring the importance of the GADD45b- ovarian carcinogenesis recapitulated the tumor-suppressive effect þ mediated function, the depletion of CD8 T cells completed of complete TME-based Gadd45b ablation (Figs. 5F and G). D D abrogated any effect of myeloid-restricted Gadd45b loss on onco- Together with our findings with Gadd45b M/ M mice, these results genesis. Indeed, GADD45b blockade appeared to also enhance demonstrate that GADD45b orchestrates the tumor-promoting the ability of the TME to support T-cell effector functions by activity of the TME across multiple cancer types by operating upregulating multiple inflammatory mediators, proinflamma- in myeloid cells. Collectively, our findings provide a mechanism tory TAM functionalities, and MHC-II molecules. Underscoring for the immunosuppressive activity of the TME that restricts þ the coordinated nature of its multiple effects on the immune tumor-based inflammation and CD8 T-cell trafficking into system, GADD45b ablation further imparted anatomic organiza- tumors and underscore the general significance of the NF-kB– tion to adaptive immune reactions, enabling them to assemble dependent mechanisms mediated by GADD45b in human into well-configured TLSs within endogenously arisen tumors. malignant disease.

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1289

Verzella et al.

Figure 7. Suppression of oncogenesis by macrophage-specific Gadd45b deletion. A, The targeting strategy for generating the loxP-flanked Gadd45b allele (Gadd45bF). B, PCR detecting the wild-type (Gadd45bþ), Gadd45bF, or excised (Gadd45bD) Gadd45b alleles in the indicated tissues from Gadd45bþ/þ (þ/þ), Gadd45bþ/F (þ/F), Gadd45bF/F (F/F), or Gadd45bF/F;LysM-cre (DM/DM) mice. C, qRT-PCR showing the relative Gadd45b mRNA levels in the indicated tissues isolated from Gadd45bF/F (F/F)orGadd45bF/F;LysM-cre (DM/DM) mice. D, Volumes of subcutaneous MCA-203 fibrosarcoma allografts in Gadd45bDM/DM (n ¼ 9) and Gadd45bF/F (n ¼ 8) mice at the times shown. E, Images of representative tumors from D at day 21. F, Summary of the cell injection schedule used in G and H. G, Relative luminescence units (RLU) of intraperitoneal ID8-Luc ovarian tumor allografts in Gadd45bþ/þ mice at the times indicated after intraperitoneal injection of Gadd45bþ/þ (n ¼ 9) or Gadd45b/ (n ¼ 12) BMDM pools. H, Bioluminescence images of representative mice from G presented as a pseudocolor scale as in Fig. 5G. I, Schematic representation of the oncogenic function of GADD45b in macrophages. C, D, and G, Values, means SEM. , P < 0.05; , P < 0.01. Also, see Supplementary Fig. S7.

1290 Cancer Res; 78(5) March 1, 2018 Cancer Research

Breaking Cancer Immunosuppression via GADD45b Deletion

Although Pikarsky and colleagues recently suggested that TLSs immune checkpoint molecules, for example, IDO and PD-1 facilitate HCC initiation and early promotion, their study inves- ligands, innate immunotherapies targeting GADD45b could tigated TLS formation in the context of the preneoplastic hepatic conceivably overcome primary resistance to adaptive immu- parenchyma and inflamed neoplastic liver outside HCCs (47). notherapies, and thereby increase response rates in currently Indeed, T- and B-cell immunity has been previously shown to recalcitrant cancer patient subsets (3, 5). counter DEN-induced HCC development in mouse models (47, Further studies will clarify the clinical benefitofcombining 48), and intratumoral TLSs have been shown to portend a GADD45b inhibitors with conventional immunotherapies. favorable clinical outcome in the majority of human cancers, Notwithstanding, the remarkable consistency of the effects of including HCC (37, 49). Future studies will determine the precise myeloid-specific Gadd45b deletion on oncogenesis and the mechanisms by which GADD45b curbs TME-based TLS forma- widespread correlation of GADD45B expression with aggressive þ tion and cytotoxic CD8 T-cell trafficking and activation, as well disease pathology across most human cancer types underscore as how these mechanisms relate to the GADD45b-mediated anti- the general clinical significance of the GADD45b-mediated inflammatory function in myeloid cells. Irrespective of the mechanism and, consequently, the potential of GADD45b- mechanisms, the profound effects of Gadd45b loss on macro- targeting therapeutics in human malignant disease. The new phage and T-cell accumulation in the TME and intratumoral TLS added focus provided by our current findings of a role of formation underscore the broad clinical potential of the GADD45b in innate immune regulation will also invigorate GADD45b-dependent anti-inflammatory axis of the IKK/NF-kB the drug discovery effort to develop novel GADD45b-targeting pathway in anticancer immunotherapy. immunotherapeutics for safely and effectively treating onco- Because GADD45b also serves as an essential NF-kB– logical diseases. regulated inhibitor of cancer cell apoptosis (24), our finding of its unexpected function in the innate immune system Disclosure of Potential Conflicts of Interest demonstrates that a single pivotal axis of the IKK/NF-kBpath- No potential conflicts of interest were disclosed. way, that is, the GADD45b-dependent axis, integrates the oncogenic mechanism suppressing tumor cell apoptosis with Authors' Contributions the mechanism restraining TME-based inflammation and T-cell Conception and design: D. Verzella, J. Bennett, M. Fischietti, A.K. Thotakura, infiltration.Hence,GADD45b serves as a central downstream E. Alesse, F. Zazzeroni, G. Franzoso Development of methodology: D. Verzella, J. Bennett, M. Fischietti, hub in the IKK/NF-kB pathway linking cancer and inflamma- A.K. Thotakura, C. Recordati, L. Tornatore, F. Zazzeroni, G. Franzoso tion. Surprisingly, GADD45b appears to perform these onco- Acquisition of data (provided animals, acquired and managed patients, genic functions via separable, tissue-specific mechanisms, provided facilities, etc.): D. Verzella, J. Bennett, M. Fischietti, A.K. Thotakura, attenuating proinflammatory p38 signaling within TAMs to F. Pasqualini, D. Capece, D. Vecchiotti, B. Di Francesco, F. Begalli, L. Tornatore, restrain proinflammatory activation, while inhibiting apoptosis S. Papa, S.J. Forbes, F. Zazzeroni of cancer cells by suppressing JNK/MKK7 signaling (Fig. 7I; Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Bennett, A.K. Thotakura, C. Recordati, D. D'Andrea, ref. 24). Future studies will clarify the precise mechanism(s) by M. De Maglie, S. Papa, A. Sica, E. Alesse, F. Zazzeroni, G. Franzoso which GADD45b curbs p38 activation in myeloid cells and Writing, review, and/or revision of the manuscript: D. Verzella, J. Bennett, confirm the role of this mechanism in oncogenesis. Irrespective M. Fischietti, A.K. Thotakura, C. Recordati, D. Capece, D. D'Andrea, F. Begalli, of these mechanism(s), our results postulate that targeting the S. Papa, S.J. Forbes, E. Alesse, F. Zazzeroni, G. Franzoso IKKb/NF-kB pathway through GADD45b inhibition provides a Administrative, technical, or material support (i.e., reporting or organizing candidate therapeutic route to counter oncogenesis by reversing data, constructing databases): J. Bennett, F. Pasqualini, T. Lawrence, A. Sica Study supervision: F. Zazzeroni, G. Franzoso TME-mediated immunosuppression and, at the same time, inducing apoptosis of cancer cells, thereby affording dual Acknowledgments fi clinical bene t. Indeed, a combination of oncogenic mechan- The work was supported in part by Cancer Research UK programme grant isms mediated by GADD45b in malignant and TME-based A15115, Medical Research Council (MRC) Biomedical Catalyst grant MR/ myeloid cells likely underpins the correlations between elevat- L005069/1 and Bloodwise project grant 15003 to G. Franzoso, the Associazione ed GADD45B expression and poor clinical outcome observed in Italiana per la Ricerca sul Cancro (AIRC) grants 1432 and 5172 and MIUR PRIN the large majority of human cancers. grant no. 2009EWAW4M_003 to F. Zazzeroni, MIUR FIRB grant no. RBA- To circumvent the limitations of IKK/NF-kB–targeting agents, P10A9H9 to E. Alesse, and the Associazione Italiana per la Ricerca sul Cancro (AIRC) grant 15585 to A. Sica. We thank N. van Rooijen for providing we recently adopted the strategy of therapeutically targeting a clodronate liposomes, I. Marigo and V. Bronte for providing the MCA-203 cell nonredundant, pathogenically critical axis of the NF-kBpath- line, and J. Dyson for assistance with animal studies. We also thank V. way,similarlymediatedbyGADD45b,inmultiplemyeloma, Tybulewicz, A. Leonardi, G. Melino, and J. Behmoaras for critical reading of rather than NF-kB globally (24, 50). Initial results from the first- the manuscript. D. Verzella and B. Di Francesco were supported by the L'Aquila in-human clinical study of GADD45b/MKK7–targeting thera- University PhD program in Experimental Medicine. M. Fischietti and D. peutics in patients with multiple myeloma preliminarily indi- Vecchiotti were supported by the L'Aquila University PhD program in Biotechnology. cate the clinical safety of these agents, alongside cancer-selective pharmacodynamic response (L. Tornatore and G. Franzoso, The costs of publication of this article were defrayed in part by the unpublished observations). These encouraging early clinical payment of page charges. This article must therefore be hereby marked results suggest that it may be possible to similarly target the advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate GADD45b-binding factor regulating p38 signaling in myeloid this fact. cells, to elicit antitumor inflammation. Notably, by converting þ þ tumors lacking a spontaneous CD8 T-cell infiltrate into CD8 Received June 22, 2017; revised November 20, 2017; accepted December 19, T-cell–infiltrated tumors and, concurrently, upregulating 2017; published OnlineFirst December 26, 2017.

www.aacrjournals.org Cancer Res; 78(5) March 1, 2018 1291

Verzella et al.

References 1. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 26. Goldman M, Craft B, Swatloski T, Cline M, Morozova O, Diekhans M, et al. 2011;144:646–74. The UCSC cancer genomics browser: update 2015. Nucleic Acids Res 2. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. 2015;43:D812–7. Cell 2010;140:883–99. 27. Marisa L, de Reynies A, Duval A, Selves J, Gaub MP, Vescovo L, et al. Gene 3. Joyce JA, Fearon DT. T cell exclusion, immune privilege, and the tumor expression classification of colon cancer into molecular subtypes: charac- microenvironment. Science 2015;348:74–80. terization, validation, and prognostic value. PLoS Med 2013;10:e1001453. 4. Topalian SL, Drake CG, Pardoll DM. Immune checkpoint blockade: a 28. Kim WJ, Kim EJ, Kim SK, Kim YJ, Ha YS, Jeong P, et al. Predictive value of common denominator approach to cancer therapy. Cancer Cell 2015;27: progression-related gene classifier in primary non-muscle invasive bladder 450–61. cancer. Mol. Cancer 2010;9:3. 5. Gajewski TF, Schreiber H, Fu YX. Innate and adaptive immune cells in the 29. Tothill RW, Tinker AV, George J, Brown R, Fox SB, Lade S, et al. Novel tumor microenvironment. Nat Immunol 2013;14:1014–22. molecular subtypes of serous and endometrioid ovarian cancer linked to 6. Noy R, Pollard JW. Tumor-associated macrophages: from mechanisms to clinical outcome. Clin Cancer Res 2008;14:5198–208. therapy. Immunity 2014;41:49–61. 30. Curtis C, Shah SP, Chin SF, Turashvili G, Rueda OM, Dunning MJ, et al. The 7. Biswas SK, Allavena P, Mantovani A. Tumor-associated macrophages: genomic and transcriptomic architecture of 2,000 breast tumours reveals functional diversity, clinical significance, and open questions. Semin novel subgroups. Nature 2012;486:346–52. Immunopathol 2013;35:585–600. 31. Hagemann T, Robinson SC, Schulz M, Trumper€ L, Balkwill FR, Binder C, 8. Qian BZ, Pollard JW. Macrophage diversity enhances tumor progression et al. Enhanced invasiveness of breast cancer cell lines upon co-cultivation and metastasis. Cell 2010;141:39–51. with macrophages is due to TNF-alpha dependent up-regulation of matrix 9. Ries CH, Cannarile MA, Hoves S, Benz J, Wartha K, Runza V, et al. Targeting metalloproteases. Carcinogenesis 2004;25:1543–9. tumor-associated macrophages with anti-CSF-1R antibody reveals a strat- 32. Papa S, Zazzeroni F, Fu YX, Bubici C, Alvarez K, Dean K, et al. Gadd45beta egy for cancer therapy. Cancer Cell 2014;25:846–59. promotes hepatocyte survival during liver regeneration in mice by mod- 10. Zhu Y, Knolhoff BL, Meyer MA, Nywening TM, West BL, Luo J, et al. CSF1/ ulating JNK signaling. J Clin Invest 2008;118:1911–23. CSF1R blockade reprograms tumor-infiltrating macrophages and improves 33. Maeda S, Kamata H, Luo JL, Leffert H, Karin M. IKKb couples hepatocyte response to T-cell checkpoint immunotherapy in pancreatic cancer models. death to cytokine-driven compensatory proliferation that promotes chem- Cancer Res 2014;74:5057–69. ical hepatocarcinogenesis. Cell 2005;121:977–90. 11. StrachanDC,RuffellB,OeiY,BissellMJ,CoussensLM,DanielD.CSF1R 34. Darani HY, Shirzad H, Mansoori F, Zabardast N, Mahmoodzadeh M. inhibition delays cervical and mammary tumor growth in murine Effects of Toxoplasma gondii and Toxocara canis antigens on WEHI-164 models by attenuating the turnover of tumor-associated macrophages fibrosarcoma growth in a mouse model. Korean J Parasitol 2009;47:175–7. and enhancing infiltration by CD8þ T cells. Oncoimmunology 2013;2: 35. Van Rooijen N, Sanders A. Kupffer cell depletion by liposome-delivered e26968. drugs: comparative activity of intracellular clodronate, propamidine, and 12. Mitchem JB, Brennan DJ, Knolhoff BL, Belt BA, Zhu Y, Sanford DE, et al. ethylenediaminetetraacetic acid. Hepatology 1996;23:1239–43. Targeting tumor-infiltrating macrophages decreases tumor-initiating cells, 36. Global Burden of Disease Cancer Collaboration. The Global Burden of relieves immunosuppression, and improves chemotherapeutic responses. Cancer 2013. JAMA Oncol 2015;1:505–27. Cancer Res 2013;73:1128–41. 37. Dieu-Nosjean MC, Goc J, Giraldo NA, Sautes-Fridman C, Fridman WH. 13. DeNardo DG, Brennan DJ, Rexhepaj E, Ruffell B, Shiao SL, Madden SF, et al. Tertiary lymphoid structures in cancer and beyond. Trends Immunol Leukocyte complexity predicts breast cancer survival and functionally 2014;35:571–80. regulates response to chemotherapy. Cancer Discov 2011;1:54–67. 38. Ley K, Pramod AB, Croft M, Ravichandran KS, Ting JP. How Mouse 14. Sica A, Mantovani A. Macrophage plasticity and polarization: in vivo Macrophages Sense What Is Going On. Front Immunol 2016;7:204. veritas. J Clin Invest 2012;122:787–95. 39. Sakurai T, Maeda S, Chang L, Karin M. Loss of hepatic NF-kappa B activity 15. Lawrence T. Macrophages and NF-kB in cancer. Curr Topics Microbiol enhances chemical hepatocarcinogenesis through sustained c-Jun N-ter- 2011;349:171–84. minal kinase 1 activation. Proc Natl Acad Sci U S A 2006;103:10544–51. 16. Ben-Neriah Y, Karin M. Inflammation meets cancer, with NF-kBasthe 40. He G, Yu GY, Temkin V, Ogata H, Kuntzen C, Sakurai T, et al. Hepatocyte matchmaker. Nat Immunol 2011;12:715–23. IKKbeta/NF-kappaB inhibits tumor promotion and progression by pre- 17. DiDonato JA, Mercurio F, Karin M. NF-kB and the link between inflam- venting oxidative stress-driven STAT3 activation. Cancer Cell 2010;17: mation and cancer. Immunol Rev 2012;246:379–400. 286–97. 18. Hagemann T, Lawrence T, McNeish I, Charles KA, Kulbe H, Thompson RG, 41. Arthur JS, Ley SC. Mitogen-activated protein kinases in innate immunity. et al. "Re-educating" tumor-associated macrophages by targeting NF-kB. Nat Rev Immunol 2013;13:679–92. J Exp Med 2008;205:1261–8. 42. Yang Y, Kim SC, Yu T, Yi YS, Rhee MH, Sung GH, et al. Functional roles of 19. Saccani A, Schioppa T, Porta C, Biswas SK, Nebuloni M, Vago L, et al. p50 p38 mitogen-activated protein kinase in macrophage-mediated inflam- nuclear factor-kappaB overexpression in tumor-associated macrophages matory responses. Mediators Inflamm 2014;2014:352371. inhibits M1 inflammatory responses and antitumor resistance. Cancer Res 43. Gaillard SL, Secord AA, Monk B. The role of immune checkpoint inhibition 2006;66:11432–40. in the treatment of ovarian cancer. Gynecol Oncol Res Pract 2016;3:11. 20. Lawrence T, Fong C. The resolution of inflammation: anti-inflammatory 44. Harding JJ, Dika El, Abou-Alfa GK. Immunotherapy in hepatocellular roles for NF-kappaB. Int J Biochem Cell Biol 2010;42:519–23. carcinoma: Primed to make a difference? Cancer 2016;122:367–77. 21. Lawrence T, Bebien M, Liu GY, Nizet V, Karin M. IKKalpha limits macro- 45. Roberts SS, Chou AJ, Cheung NK. Immunotherapy of Childhood Sarco- phage NF-kappaB activation and contributes to the resolution of inflam- mas. Front Oncol 2016;5:181. mation. Nature 2005;434:1138–43. 46. Fearon DT. Immune-suppressing cellular elements of the tumor microen- 22. Papa S, Zazzeroni F, Bubici C, Jayawardena S, Alvarez K, Matsuda S, et al. vironment. Annu Rev Cancer Biol 2016;1:13.1–13.15. Gadd45 beta mediates the NF-kappa B suppression of JNK signalling by 47. Finkin S, Yuan D, Stein I, Taniguchi K, Weber A, Unger K, et al. Ectopic targeting MKK7/JNKK2. Nat. Cell Biol 2004;6:146–53. lymphoid structures function as microniches for tumor progenitor cells in 23. De Smaele E, Zazzeroni F, Papa S, Nguyen DU, Jin R, Jones J, et al. Induction hepatocellular carcinoma. Nat Immunol 2015;16:1235–44. of gadd45beta by NF-kappaB downregulates pro-apoptotic JNK signalling. 48. Schneider C, Teufel A, Yevsa T, Staib F, Hohmeyer A, Walenda G, et al. Nature 2001;414:308–13. Adaptive immunity suppresses formation and progression of diethylni- 24. Tornatore L, Sandomenico A, Raimondo D, Low C, Rocci A, Tralau-Stewart trosamine-induced liver cancer. Gut 2012;61:1733–43. C, et al. Cancer-selective targeting of the NF-kB survival pathway with 49. Wada Y, Nakashima O, Kutami R, Yamamoto O, Kojiro M. Clinicopatho- GADD45b/MKK7 inhibitors. Cancer Cell 2014:26:495–508. logical study on hepatocellular carcinoma with lymphocytic infiltration. 25. Tomczak K, Czerwinska P, Wiznerowicz M. The Cancer Genome Atlas Hepatology 1998;27:407–14. (TCGA): an immeasurable source of knowledge. Contemp Oncol 2015;19: 50. Karin M.Whipping NF-kB to submission via GADD45 and MKK7. Cancer A68–A77. Cell 2014;26:447–49.

1292 Cancer Res; 78(5) March 1, 2018 Cancer Research

GADD45β Loss Ablates Innate Immunosuppression in Cancer

Daniela Verzella, Jason Bennett, Mariafausta Fischietti, et al.

Cancer Res 2018;78:1275-1292. Published OnlineFirst December 26, 2017.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-17-1833

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2017/12/23/0008-5472.CAN-17-1833.DC1

Cited articles This article cites 50 articles, 9 of which you can access for free at: http://cancerres.aacrjournals.org/content/78/5/1275.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/78/5/1275.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/78/5/1275. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.